15:00:43 #startmeeting neutron_l3 15:00:44 Meeting started Thu Jan 19 15:00:43 2017 UTC and is due to finish in 60 minutes. The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:47 The meeting name has been set to 'neutron_l3' 15:00:50 \o/ 15:00:58 hi 15:02:46 Agenda for today is here: 15:02:57 #link https://etherpad.openstack.org/p/neutron-l3-subteam 15:03:26 #topic Annoucements 15:04:00 First, Ocata-3 is next week. We made it to the end of this short cycle 15:04:09 Time sure does fly 15:05:03 The PTG on February 24th is approaching fast 15:05:43 I'm still waiting for my travel to be approved.. Not sure if I'll make it 15:05:54 82 spots remain open 15:06:05 as of 30 seconds ago 15:06:27 john-davidge: hopefully you'll get approved. Looking forward to see you there 15:06:49 mlavalle: You too, keeping my fingers crossed :) 15:07:46 Any other annoucements? 15:08:59 ok, moving on 15:09:05 #topic Bugs 15:09:57 First one is https://bugs.launchpad.net/neutron/+bug/1610483 15:09:57 Launchpad bug 1610483 in neutron "Pluggable IPAM rollback mechanism is not robust" [High,In progress] - Assigned to Aliaksandr Dziarkach (aliaksandr-dziarkach) 15:10:10 This bug has a solution proposed here: 15:10:25 #link https://review.openstack.org/390594 15:10:51 It is in my list of things to review. Hasn't gotten there yet :-( 15:11:14 Encourage others to take a look 15:11:40 Any comments? 15:11:42 i reviewed but wasn't super familiar with the sql stuff, so i added kevin benton tot he review 15:11:54 haleyb: great. thanks! 15:12:09 it seemed good though 15:12:19 yeah, I have the same impression 15:13:04 ok, next one is https://bugs.launchpad.net/neutron/+bug/1570122 15:13:04 Launchpad bug 1570122 in neutron "ipv6 prefix delegated subnets are not accessable external of the router they are attached." [High,In progress] - Assigned to John Davidge (john-davidge) 15:13:37 https://review.openstack.org/#/c/407025/ 15:13:47 mlavalle: So haleyb has raised a good point in his latest review 15:13:48 yeap, you reviewed it 15:14:05 mlavalle: Going to need to jump down the rabbit hole a bit to fix it 15:14:22 mlavalle: I expect to have a newer, better, shinier patch by the end of tomorrow 15:14:32 john-davidge: Thanks! 15:15:21 Next one is https://bugs.launchpad.net/neutron/+bug/1627424 15:15:21 Launchpad bug 1627424 in neutron "FlushError on IPAllocation" [High,Confirmed] - Assigned to Miguel Lavalle (minsel) 15:15:49 I've continued looking at this one and have noticed a few things that have drawn my attention: 15:17:23 1) when the problem arises as a consequence of deleting subnets, it is triggered by the db plugin when removing ips from ports. That leads to this exception being caught: https://github.com/openstack/neutron/blob/master/neutron/db/db_base_plugin_v2.py#L1259 15:18:08 Notice the comment about port update and subnet delete interleaving. Does anyone have further insight? 15:19:13 mlavalle: looks like armax wrote it 15:19:36 john-davidge: that's actually good insight... I'll ping him 15:20:01 mlavalle: git blame to the rescue :) 15:20:18 yeap, I should have thought of that :-) 15:22:16 2) That exception caught by the db plugin is raised here: https://github.com/openstack/neutron/blob/master/neutron/ipam/drivers/neutrondb_ipam/driver.py#L223. It is raised when this query delete return 0 rows: https://github.com/openstack/neutron/blob/master/neutron/ipam/drivers/neutrondb_ipam/db_api.py#L142 15:23:28 Reading the sqlalchemy docs http://docs.sqlalchemy.org/en/latest/orm/query.html#sqlalchemy.orm.query.Query.delete, it seems query deletes operate outside the normal transaction / session protection 15:23:44 that might be the cause of the flush error. Any comments? 15:26:33 If there are no comments, I'll continue digging :-) 15:26:34 Next one is https://bugs.launchpad.net/neutron/+bug/1627480 15:26:34 Launchpad bug 1627480 in neutron "create_port can succeed without returning fixed_ips on all requested subnets" [High,Confirmed] 15:26:34 It doesn't have an owner yet. Looking for a volunteer :-) 15:26:47 Any takers / comments? 15:27:32 ok, next one is https://bugs.launchpad.net/neutron/+bug/1509004 15:27:33 Launchpad bug 1509004 in neutron ""test_dualnet_dhcp6_stateless_from_os" failures seen in the gate" [High,Confirmed] 15:27:56 mlavalle: On that note, I have to run! Sorry all, see you next week :) 15:28:13 This one returned last week. haleyb was keeping an eye on it. Any updates? 15:28:18 mlavalle: I guess i don't know the history of that one, since i see there was a fix ? 15:28:27 sorry, slow response 15:29:50 haleyb: going back to the previous one, you are right there was a proposed fix 15:29:53 mlavalle: no new info on the dhcpv6 bug though, ^^comment was for the other 15:30:16 haleyb: again, referring to the previous one, I'll dig a little 15:30:27 and see if that takes care of it 15:30:36 mlavalle: ok, as there was a change merged and backported, but i'm thinking there was still something missing 15:30:49 that is also my impression 15:31:09 but I'll try to scope the current problem a little better 15:32:17 Any other bugs from the team? 15:33:43 ok, let's move on 15:33:52 #topic Prefix Delegation 15:33:59 baoli: hi! 15:35:53 I see that baoli updated the etherpad. Let's move on 15:36:10 #topic Routed Networks 15:37:02 The rest client for the placement API patchset was approved last night by kevinbenton: https://review.openstack.org/#/c/414726 15:37:55 The second patchset, which is the piece that updates ipv4 inventories using the placement API got some comments from kevinbenton 15:38:20 last night. I'll address them today and push an update 15:38:35 Thanks to haleyb for the thorough reviews :-) 15:38:37 yes, good comments, will review when i see it 15:39:45 I don't see tidwellr, so we will skip BGP Dynamic Routing 15:39:51 #topic Open Agenda 15:40:09 Any other items we need to discuss today? 15:41:07 i don't have any 15:41:41 ok, thanks for attending 15:41:50 Enjoy the rest of your days! 15:41:54 #endmeeting